Information Maximizing Optimal Sensor Placement Robust Against Variations of Traffic Demand Based on Importance of Nodes

Jordan Ivanchev, Heiko Aydt, Alois C. Knoll

Abstract: This paper defines a measure of importance for a node (intersection) in a transportation network, based on its topology and traffic demand. Consequentially, the measure is used in order to solve the sensor placement problem by maximizing the information gain in terms of users' routing choices by sensing the most uncertain areas in the system. It is demonstrated that utilizing the proposed strategy makes the performance robust against short- and long-term variations of traffic patterns. Finally, a method for finding the optimal number of sensors to be installed in a city is proposed. It models and maximizes the utility stemming from the tradeoff between cost, performance, robustness, and reliability of the sensor placement problem solution.